On execution models, honest pricing, and connecting the tools operators actually run on.
One run is one trigger, however many steps it has. Plans differ by how many runs a month, how many people and how much support. No overage, no surprise bill.
READ →n8n runs downstream once with the whole array. Make runs it once per item. OpFlow does both — set per bot, override per node.
Point Claude — or any MCP client — at OpFlow and it can list, read, run, and manage your bots. Here's how, and why.
Retries, Try/Catch, dedicated error branches, and replay — on every account, because real work fails sometimes.
Pin test data, run from any node, read every step's payload. Building an automation should feel like debugging code — because it is.
OpFlow lives inside the platform that emits the events — so a form submission or a metric crossing a threshold can start a bot with zero webhook plumbing.